Diminishing Returns of Project Size

Engineers are an interesting lot. As the jokes go, we find the talking frogs more interesting than the princesses that they claim to be. Best of all is the glass that is half-full or half empty; to the engineer the glass is twice the size it needs to be.

Engineering students take a class in a strange form of Economics called Engineering Economics. Along with some history, including the work of Fredrick Taylor, Frank and Lilian Gilbreth, and Henri Fayol, this class introduces the engineering student to the concept of the Law of Diminishing Returns. To be fair, economics students get a full dose of the Law of Diminishing Returns, along with Economies of Scale, Constant Returns to Scale, and Diseconomies of Scale.

I am thinking about this when I consider how to illustrate to a client how bigger orders do not always generate lower per-unit costs. The client, influenced by a sourcing consultant, thinks that bigger projects should always cost less per unit because of the Economies of Scale. The consultant’s influence is strong because the consultant has demonstrated to the client that they can achieve greater first-cost reductions by consolidating purchases from multiple suppliers and buying from a single supplier or a small group of suppliers. It is early enough in the process that the ugly realities of internal resistance, excess inventory, and poor service have yet to emerge. The internal resistance is already building, so it will not be long before the rest of the problems emerge.

Bigger is not always better

I believe that there is a perfect size for every application. We understand this with clothing, with the basic four-segment spread of Small, Medium, Large and Extra Large. Sometimes the fit is more important, as with a man’s suit jacket or a woman’s dress, so the size chart expands. Statisticians apply the normal distribution, the Bell Curve, to define where demand or supply may fall in the perfect world. The more refined the need, the more effort we put into sizing the garment to the person.

What is the size of the perfect order? What do we even mean when we say Perfect Order? Is there a perfect order size, or is it really more the measure of the Optimized Order Size? Optimal is an interesting concept that is misunderstood. The graph of the Law of Diminishing Returns above is a picture of the concept of optimal. I say people misunderstand the concept of optimal because they tend to go past the point of optimum, reacting to how things are getting worse as they try harder. Top performers are much more sensitive to changes in results and required effort. That is why top performers will stop adding more effort, money, or resources to a project before the output reaches the point of maximization, the top of the curve.

Notice I said that that the top of the curve is the point of maximization, where we are getting the absolute maximum out of the system, the point at which, if we add one more input, the output actually drops. The optimal point is the inflection point before the top of the curve, the point where the rate in improvement decreases with each additional input. High performers stop at the optimum point, understanding that more actually starts to create less.

I have a client who is looking to build a new distribution center. The client bought the land to build a million-square-foot building, with the associates trailer storage and parking lots. The architect and the construction company are telling the client that it is better to build the entire building than only the first 400,000 square feet, because the cost to build the full million will be less per square foot (pfs) than to build only the four hundred thousand. The client asked me if this was true. Sure enough, it will cost only $37 pfs to build the million and $41 to build the 400,000.

Then I asked the client how much cash they wanted to tie up in a building they could not use. The one-million square foot building requires $37 million to build, but the 400,000 square foot building costs only $16.4 million. I asked the client to think of what else they could do with the $20.6 million the extra space would cost. Before they answered, I asked them to think about whether the cost of the smaller building was really an investment in the future expansion of the building. In this case, the issue was not the Law of Diminishing Returns, but really the concept of utility; what else could you use that money for that could create a higher return than 600,000 square feet of unused building?

When does leverage stop?

Working with the Theory of Constraints (TOC), the challenge is how to get more revenue out of a system once it reaches its capacity constraints. The Law of Diminishing Returns plays a role in TOC thinking. To push the constraint to work at maximum utility, managers put in more effort to monitor, measure, and plan the flow through the constraint. TOC helps steepen the curve between the blue X in the graph above and the top of the curve. Additional management effort does not always mean more cost to the system, unless management is the constraint. Break the constraint by adding another machine, more people, or more space, and you may overload the capacity of the management in place.

That is the problem with projects. While there are many different constraints to any project, the key constraint that most people overlook is the management of the project. Projects by default are abnormal events. Projects are not repetitive in the sense of place, personnel, or scope. Yes, a WMS system may be the same from DC to DC in the same company, but other factors change, which changes the outcomes. The same is true for construction projects to install equipment and machines in an operation. The variance from project to project can erode the capacity of the Project Managers to execute the needed planning, oversight, and direction.

In a review of a giant shelving project, the end user asked why the project was behind schedule. Building construction and materials delivery matched the schedule. As the project slipped, the installers brought more labor to the site. Still the project slipped. None of these was the constraint; the project manager was. On-site for three days a week, the project manager spent over half of his time in meetings about the slipping schedules, and writing reports as required by the contracts. The limited time he spent on site was the root cause, because of the contractual requirement that the installation crews could not move on to the next phase of work until after the Project Manager had inspected and approved the work. More labor did not solve the problem, because as they completed the work faster, they waited longer for the Project Manager to inspect the work. The Project Manager inspected work completed on Thursday 2nd shift the following Tuesday afternoon when he returned. Lifting the "inspection before moving on" requirement removed this constraint, and the project moved ahead and made up for the lost time.

Economy of Scale always encounters resistance. Wind resistance doubles with the speed or size of the vehicle moving through the air. Project costs escalate as the project grows larger or as the deadline moves up. Just as we can change the shape of the car so that it moves through the air with less resistance, we can change the way we plan and manage projects to overcome the resistance of size and speed. However, that effect of that change goes only so far before we hit the top of the curve and we see progress slow down.